Deutsch Intern
    Data Science Chair

    Our Paper "Do Different Deep Metric Learning Losses Lead to Similar Learned Features?" is currently trending on DeepAI.org

    05/11/2022

    Our recent ICCV paper "Do Different Deep Metric Learning Losses Lead to Similar Learned Features?" is currently trending on the AI research discovery platform deepai.org.

    The paper, that was published at ICCV 2021, investigates the influence of input features on deep metric learning models. We found that even though different deep metric learning loss function perform quite similarly when compared fairly, there are differences in learned features between different types of loss function. We performed low-level (pixel) and high-level (attributes) analyses and introduced new analysis methods to discover features that influence deep metric learning methods, paving the way for more interpretable and better deep metric learning research.

    Currently, our paper trends on https://deepai.org, an AI research discovery platform. This website collects papers, datasets, features lexicons of AI jargon, and helps with keeping up to date with AI news. We are happy that our paper gets so much attention on this platform.

    The paper can be found here: https://deepai.org/publication/do-different-deep-metric-learning-losses-lead-to-similar-learned-features

    The code and data for the paper can be found here: https://github.com/LSX-UniWue/DML-analysis

    Back